https://skillfed.io/news/2026-09-23/1-of-tokens-can-be-enough-on-gradient-estimation-in-on-policy-distillation SkillFed en 2026-09-23T14:15:09+00:00 1% of Tokens Can Be Enough: On Gradient Estimation in On-Policy Distillation https://skillfed.io/news/2026-09-23/anyjev SkillFed en 2026-09-23T14:15:09+00:00 nokia-applied-research/AnyJev https://skillfed.io/news/2026-09-23/care-experience-guided-atomic-corrective-execution-for-vision-language-action SkillFed en 2026-09-23T14:15:09+00:00 CARE: Experience-Guided Atomic Corrective Execution for Vision-Language-Action Policies https://skillfed.io/news/2026-09-23/deep-persona-a-psychologically-grounded-architecture-and-evaluation-framework SkillFed en 2026-09-23T14:15:09+00:00 Deep Persona: A Psychologically Grounded Architecture and Evaluation Framework for Role-Playing Agents and Simulations https://skillfed.io/news/2026-09-23/gamehorizon-suite-multi-horizon-data-and-evaluation-in-gameplay SkillFed en 2026-09-23T14:15:09+00:00 GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay https://skillfed.io/news/2026-09-23/geo-sleuth SkillFed en 2026-09-23T14:15:09+00:00 Oldcircle/geo-sleuth https://skillfed.io/news/2026-09-23/harness-zero-harness-distillation-via-agent-as-harness SkillFed en 2026-09-23T14:15:09+00:00 Harness-Zero: Harness Distillation via Agent-as-Harness https://skillfed.io/news/2026-09-23/jev-align SkillFed en 2026-09-23T14:15:09+00:00 sutro-sh/jev-align https://skillfed.io/news/2026-09-23/jev-browser-use SkillFed en 2026-09-23T14:15:09+00:00 wy-coliney/jev-browser-use https://skillfed.io/news/2026-09-23/jev-mcp SkillFed en 2026-09-23T14:15:09+00:00 jkudish/jev-mcp https://skillfed.io/news/2026-09-23/jev-router SkillFed en 2026-09-23T14:15:09+00:00 gargpratyush/jev-router https://skillfed.io/news/2026-09-23/nullmotion SkillFed en 2026-09-23T14:15:09+00:00 blixvip/NullMotion https://skillfed.io/news/2026-09-23/one-to-more-more-to-one-category-aware-iterative-expert-training-for-software SkillFed en 2026-09-23T14:15:09+00:00 One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agents https://skillfed.io/news/2026-09-23/onpanda-efficient-annotation-of-on-policy-alignment-data-for-llms-and-agents SkillFed en 2026-09-23T14:15:09+00:00 onPanda: Efficient Annotation of On-Policy Alignment Data for LLMs and Agents via Token-Level Correction https://skillfed.io/news/2026-09-23/opendlss-nr SkillFed en 2026-09-23T14:15:09+00:00 maanHimself/OpenDLSS-NR https://skillfed.io/news/2026-09-23/realtime-venus-a-full-duplex-interaction-system-with-asynchronous-delegation SkillFed en 2026-09-23T14:15:09+00:00 Realtime-Venus: A full-duplex interaction system with asynchronous delegation https://skillfed.io/news/2026-09-23/rizzo-flow SkillFed en 2026-09-23T14:15:09+00:00 Rizzo-AI-Academy/rizzo-flow https://skillfed.io/news/2026-09-23/rrsi-regularized-recursive-self-improvement-of-agent-harnesses SkillFed en 2026-09-23T14:15:09+00:00 RRSI: Regularized Recursive Self-Improvement of Agent Harnesses https://skillfed.io/news/2026-09-24/3dicon SkillFed en 2026-09-24T14:15:13+00:00 samyost1/3dicon https://skillfed.io/news/2026-09-24/agent-jev SkillFed en 2026-09-24T14:15:13+00:00 malevrigns/agent-jev https://skillfed.io/news/2026-09-24/all-in-one-multilingual-scene-text-recognition-with-script-aware-mixture-of SkillFed en 2026-09-24T14:15:13+00:00 All-in-One Multilingual Scene Text Recognition with Script-aware Mixture-of-Experts https://skillfed.io/news/2026-09-24/emergent-collusion-in-long-horizon-llm-agent-interaction SkillFed en 2026-09-24T14:15:13+00:00 Emergent Collusion in Long-Horizon LLM Agent Interaction https://skillfed.io/news/2026-09-24/flash-dllm-io-aware-kv-caching-and-parallel-decoding-for-fast-memory-efficient SkillFed en 2026-09-24T14:15:13+00:00 Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs https://skillfed.io/news/2026-09-24/gae-learning-a-geometry-native-latent-space-for-3d-consistent-world-generation SkillFed en 2026-09-24T14:15:13+00:00 GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation https://skillfed.io/news/2026-09-24/jev-as-a-judge-accept-when-confident-escalate-when-unsure SkillFed en 2026-09-24T14:15:13+00:00 JEV-as-a-Judge: Accept When Confident, Escalate When Unsure https://skillfed.io/news/2026-09-24/jev-codex-router SkillFed en 2026-09-24T14:15:13+00:00 0xNatoshi/jev-codex-router https://skillfed.io/news/2026-09-24/jev-visual SkillFed en 2026-09-24T14:15:13+00:00 hr98w/jev-visual https://skillfed.io/news/2026-09-24/lean-pool-an-ai-maintained-archive-of-formalized-mathematics SkillFed en 2026-09-24T14:15:13+00:00 Lean Pool: An AI-Maintained Archive of Formalized Mathematics https://skillfed.io/news/2026-09-24/llm2jev SkillFed en 2026-09-24T14:15:13+00:00 Yinsongxu/LLM2Jev https://skillfed.io/news/2026-09-24/magpie SkillFed en 2026-09-24T14:15:13+00:00 yetone/magpie https://skillfed.io/news/2026-09-24/ovis-embedding-pushing-the-frontiers-of-universal-omni-modal-embeddings SkillFed en 2026-09-24T14:15:13+00:00 Ovis-Embedding: Pushing the Frontiers of Universal Omni-Modal Embeddings https://skillfed.io/news/2026-09-24/pdoomvideo SkillFed en 2026-09-24T14:15:13+00:00 JohnHeibel/PDoomVideo https://skillfed.io/news/2026-09-24/procedural-film SkillFed en 2026-09-24T14:15:13+00:00 kuhnhomeuk-cell/procedural-film https://skillfed.io/news/2026-09-24/recursive-self-improvement-of-ai-research-agents SkillFed en 2026-09-24T14:15:13+00:00 Recursive self-improvement of AI research agents https://skillfed.io/news/2026-09-24/stablevq-practical-guidelines-for-stable-vector-quantized-tokenizer-training SkillFed en 2026-09-24T14:15:13+00:00 StableVQ: Practical Guidelines for Stable Vector-Quantized Tokenizer Training https://skillfed.io/news/2026-09-24/typesafe-mcp SkillFed en 2026-09-24T14:15:13+00:00 itsmostafa/typesafe-mcp